Field of the Invention
The invention described herein relates to marketing, and in particular relates to analysis of consumer behavior.
Related Art
Any seller of goods or services seeks to expand its consumer base. This can be done with any one of several marketing tactics. These tactics can include, for example, traditional advertising, or the offering of incentives to purchase. Such incentives may include rebates or discounts. Marketing tactics are most efficient when directed at a particular subset of potential customers, i.e., those consumers who are most likely to be interested in the product or service. Marketing tactics may also be directed at existing customers, in an effort to increase their spending. It is therefore prudent to identify those consumers, either existing customers or potential customers, who are most likely to want or need a particular product or service.
One method by which such consumers can be identified is collection of information about consumers, to see which ones have one or more characteristics that identify a consumer as a likely buyer. A specific characteristic can be modeled as a point on a continuum. Such a continuum 100 is shown in FIG. 1. Here, a particular credit rating, for example, can be modeled as a point on a spectrum that represents all possible credit ratings or categories thereof. A consumer 105 has a rating of 470, which is shown as the corresponding point on continuum 100. It may be convenient for analytical purposes to organize the points of such a continuum into categories. These are shown as categories 110, 120, and 130. A consumer will therefore be mapped to a particular category based on his or her credit rating. Consumer 105 belongs in category 120 of continuum 100 in the illustration.
Such a spectrum or continuum can be viewed as a dimension or attribute. Credit rating is one example of a dimension; another dimension might be income level; another might be age. If it is believed that prospective buyers have a certain characteristic, the targeted consumers will be those that are clustered around a certain point or that fall into a particular category with respect to the appropriate attribute or dimension. A bank may wish to offer a credit card for persons having a credit rating in category 120 of FIG. 1, for example. In this example, the bank would identify people having such a credit rating, and send such people mailings that describe the terms and benefits of a card offered by the bank.
In another example, an automobile manufacturer may have a model which the manufacturer wishes to market to older adults. The automobile manufacturer may identify consumers who are 55 years old, plus or minus some value, e.g., consumers between the ages of 50 and 60. Prospective customers in this demographic may then be sent advertising or rebate offers with respect to that model. Both of these examples represent single dimension analysis. In the first example, prospective customers are identified on the basis of a single attribute or dimension, i.e., credit rating. In the second example, potential borrowers are identified on the basis of a single attribute, i.e., age.
Targeting specific consumers through analysis of a single attribute, however, has shortcomings. Generally, the set of consumers identified through such a method is too broad. For example, just because a person is between the ages of 50 and 60, does not mean that such a person has any interest in purchasing a certain automobile. Generally, only a small subset of people between the ages of 50 and 60 have any interest in purchasing an automobile. Therefore, multi-dimensional analysis is sometimes used. For example, the above automobile manufacturer may choose to market only to those people who are between the ages of 50 and 60, and who have a sufficiently high credit rating that would enable the potential customer to handle a substantial car loan. This is illustrated in FIG. 2. Here, consumer 105 has a credit rating of 470, as shown with respect to the credit rating dimension. Consumer 105 is also 45 years of age, as shown with respect to the age dimension. These two characteristics of consumer 105 are therefore modeled as a single point in a two-dimensional space. In the illustration, the age dimension is organized into categories 210, 220, and 230 in a similar manner to the categorization of the credit rating dimension. Consumer 105 therefore falls into age category 230. In the context of the illustrated two-dimensional space, consumer 105 falls into two-dimensional category 280. Category 280 represents consumers who are between the ages of 40 and 50, and who have a credit rating between 400 and 500.
Additional dimensions can also be added. The automobile manufacturer, for example, may choose to market to adults who are between the ages of 40 and 50, who have a relatively high credit rating, and who have incomes greater than $150,000 a year. Generally, additional dimensions can be included in the analysis in order to narrow the set of potential customers.
While using a large number of dimensions serves to produce a narrowly focused set of consumers, performing such an analysis can be difficult. First, it may be difficult to collect enough data for potential customers. Ideally, a company may desire a variety of demographic information with respect to potential customers, along with detailed information with respect to their spending habits. All this information may difficult to obtain. In addition, analysis involving a large number of dimensions can be computationally difficult. Hundreds of attributes are available. The storage and processing of this volume of information may not be feasible or cost effective. In contrast, the use of a lower number of dimensions allows for relatively easy processing, but produces a relatively large, and unfocused set of potential consumers. This would result in the direction of marketing tactics at a set of potential customers that is too large. Many people in such a subset would have no interest in the offered product or service.
Note also that even if it were practical to handle large amounts of data thereby enabling sophisticated multi-dimensional analysis, the results may still be inadequate. This is because attributes, such as those presented above, represent snapshots of a potential customer's position. At any given point in time, a potential customer has a certain credit rating, is of a certain age, and has a certain income. Such information offers nothing with respect to trends or changes in such attributes. A person with a high income and a high credit rating may appear to be a good target for a mortgage refinancing offer. But such a person might not be as attractive a candidate if the person were about to retire and therefore experience a reduction in income. A potential customer's attributes may therefore be of interest, but trends in such attributes for the prospective customer are also of interest. Historically, such trends are not considered.
There is a need, therefore, for a method by which prospective customers can be targeted, wherein a large number of attributes can be considered in a way that allows practical and efficient data storage and processing. Moreover, such targeting also needs to include, as attributes, changes in the status of a prospective customer with respect to such attributes.